DRIP - Know which A-B test wins before you run it

If you run an online shop doing at least six figures a month, and you want to know which a B test on your site is going to win before you run it, then this is for you.

Hey.

I'm Fabi, founder of Drip, and my team and I have run over 4,000 A B test for more than 250 brands in the span of eight years.

These are brands like Snox AG1, Hornbach Kicks, Gieswein, Blackroll, and, many more.

And we have made them altogether over 500 million euros in extra revenue.

Now, most of you might already run tests, so I'll skip the basics and get straight to the one thing we can do that nobody else can.

Knowing whether a test wins before it goes live.

The problem with regular A B testing these days is simple.

Most test ideas are complete guesses coming from a competitor's shop because you've seen it in a checklist or whomever argued the loudest in the meeting.

So about one test in 10 or two tests in 10 win.

And those that test run maybe three to four tests a month, which makes about four small wins a year.

Testing only pays when the wins stack every uplift on top of the last.

And for that you, you need a lot of tests at a high win rate.

No testing tool on the market helps you with either.

It splits the traffic and it counts, and that's where it stops.

It can tell you which ideas are worth building and which are not.

And it remembers nothing from your last test.

So we build exactly that.

It's called Apex, an A B testing tool.

And the difference fits in one sentence.

Every other tool starts every test from zero.

Apex starts every test from 4.3 million.

That's how many tests we've recorded for more than 150,000 shops from huge brands, from medium brands, and also from small brands.

So when you type an idea into Apex, it already knows how that idea went for shops like yours.

Then it adds what our behavioral psychologists find in your review.

Support tickets, heat maps and session recordings.

Results in one number next to every idea, the likelihood it wins.

The best bets get built and the weak ones die on paper.

And six to ten experiments run side by side, parallel.

We call that prediction based experimentation.

Let me show you inside the tool.

Now, this is the backlog in Apex for Snox.

What you see here on are all the experiments or an assortment of experiments that are ranked for Snox.

So that means these different A B tests are planned for Snox.

They each have a win likelihood behind them.

For example, the explicit guarantee scope in USP has a win likelihood of 67.

Now let's look into one of those ideas.

The durability micro quotes on mobile plp.

We have the hypothesis and the description here and the win likelihood.

Now how does this win likelihood come to place?

Basically we have the win prediction.

So you know how many similar tests have been run across the market, social proof, etc.

And how many of them have won.

Then we have the business impact, the ease to build, how noticeable it is and the expert read and all of these come together for the Apex for the likelihood to win score.

Now let's look at those.

This test here, this is the test that we have planned here.

A sub headline on the PLP obviously.

Yeah, very present.

So high prioritization here.

Here you can add research notes, comments, etc.

That's how one A B test looks.

Now what if you want to add an idea?

We just click add an idea and describe what we want to do.

Let's say you want a free shipping progress bar in cart and then we can say okay, this is a cart.

I want this on all devices.

If you want you can use the type, the main metrics and the effort.

This was auto, as you can see.

This was auto tagged here already.

So you even don't need to do that.

And if you want you can add more stuff.

If you add that to backlog, which I will not do.

Now this will be added to your backlog right away and we will score it for you.

Now one of my favorite features, and this is a demo account for gymshark, we're not working with them, is a heat map out of the box.

If you run any A B test in Apex, you are able to see a heat map of how people have behaved in different variants.

So you can see how different variants affect your users or don't.

Now the last thing I want to show is the testing program.

Basically it's the one page and you need to look at to see how your testing program is performing in the web, the app, or last 12 months, etc.

How many wins we have, how many tests you have live and more.

That's what knowing before you build looks like.

Last quarter, 55% of our A B tests won compared to 27% in 2024 when we still did regular A B testing.

And here are some more results.

For Oceans Apart, for example, we ran 34 tests in six months and 17 of them won.

And more revenue comes in every month.

If you want that in your shop, there are two ways to work together.

One, we do it all and you watch it Earn.

That means we install the testing tool or use yours.

For now, our team researches your shop, fills the backlog, scores every idea using the prediction mechanism, designs and builds the best ones, checks them on real devices and launches them and reads results.

It's basically all taken care of.

All of that happens inside Apex, where you can watch every test Earn.

Your team doesn't have to do a thing.

Option two, your team does the testing and we guide it.

You get Apex and our experiments running so many A B tests with the same database and the same odds we use.

And your team builds and launches and checks the tests.

Every two week.

We get on a call, bring you fresh ideas and go through what the last test earned.

And because we can predict what wins, we can promise what happens.

At least 10% more revenue in under six months.

Or we keep working from month six months seven, at no extra fee until you have it.

If you're interested, book a quick call with my team.

We'll go through your shop together, see if it's a fit, and if it is, which of the two ways makes sense.

Looking forward to it.

Cheers.